- From: Champion, Mike <Mike.Champion@SoftwareAG-USA.com>
- Date: Wed, 21 May 2003 21:46:02 -0400
- To: www-ws-arch@w3.org
> -----Original Message----- > From: Ugo Corda [mailto:UCorda@SeeBeyond.com] > Sent: Wednesday, May 21, 2003 4:30 PM > To: Champion, Mike; www-ws-arch@w3.org > Subject: RE: Separate concepts for "service" and > "targetResource?" (was > RE : /service/@targetResource ?) > > > That's not my reading of the working draft. My understanding > is that the WSD proposal is for a "service" to have an > attribute that points to a URI. This URI identifies something > other than the "service" itself. Your reading of the WSDL draft is probably correct -- they don't care about the "resource" / "targetResource" other than to know its identity. WSA has to care about the provider's view of the actual implemented service (e.g., it is the thing that needs to be managed, and the thing that has semantics, and performs some actual task). Thus WSA the people in Rennes tended to want to split the service description concept of "service" from the provided actual code concept of "service." > My understanding of an agent is that it is a piece of code. > Again, if I replace that piece of code with another because > of any reason, do I get a different URI? And why should the > service care (assuming the semantics remain the same)? This is an important point. I think we just sortof waved our hands over it in Rennes. Clearly the URI does not change when an agent is recompiled, or translated from C# to Java without changing defined functionality. Does it change if the back-end process is moved from an Oracle ERP system to a SAP ERP system? Maybe, because the semantics have probably changed ... but maybe not if the exposed operations are the same from a web service requresters point of view.
Received on Wednesday, 21 May 2003 21:46:03 UTC